Michael Doyle

North Park

San Diego, CA

I am broadly interested in open source, natural language processing, applying language models to code, adversarial AI, and turning research code into production.

Specifically, my work has entailed researching, training, testing, and deploying models in both TensorFlow and PyTorch for natural language processing, adversarial computer vision and text, and acoustics. I’ve also worked on simulation and tooling for reinforcement learning.

I enjoy making code pretty and automating as much of my job as possible.

selected publications

NAACL

Testing the Effect of Code Documentation on Large Language Model Code Understanding

William Macke, and Michael Doyle

North American Association of Computational Linguistics (NAACL), 2024

arXiv Bib PDF Poster Website

@article{macke2024testing,
  title = {Testing the Effect of Code Documentation on Large Language Model Code Understanding},
  author = {Macke, William and Doyle, Michael},
  journal = {North American Association of Computational Linguistics (NAACL)},
  year = {2024},
  organization = {Association for Computational Linguistics},
}

SPIE

The vulnerability of UAVs: An adversarial machine learning perspective

Michael Doyle, Josh Harguess, Keith Manville, and 1 more author

In Geospatial Informatics XI , 2021

Bib Website

@inproceedings{doyle2021vulnerability,
  title = {The vulnerability of UAVs: An adversarial machine learning perspective},
  author = {Doyle, Michael and Harguess, Josh and Manville, Keith and Rodriguez, Mikel},
  booktitle = {Geospatial Informatics XI},
  volume = {11733},
  pages = {81--92},
  year = {2021},
  organization = {SPIE},
}

NeurIPS

Reinforcement Learning for Wildfire Mitigation in Simulated Disaster Environments

Alexander Tapley, Marissa Dotter, Michael Doyle, and 5 more authors

Neural Information Processing Systems (NeurIPS), 2023

arXiv Bib PDF Code Poster Website

@article{tapley2023reinforcement,
  title = {Reinforcement Learning for Wildfire Mitigation in Simulated Disaster Environments},
  author = {Tapley, Alexander and Dotter, Marissa and Doyle, Michael and Fennelly, Aidan and Gandikota, Dhanuj and Smith, Savanna and Threet, Michael and Welsh, Tim},
  journal = {Neural Information Processing Systems (NeurIPS)},
  year = {2023},
}

AMTA

Practical Attacks on Machine Translation Using Paraphrase

Elizabeth M. Merkhofer, John Henderson, Abigail S. Gertner, and 2 more authors

In Proceedings of the 15th biennial conference of the Association for Machine Translation in the Americas (Volume 1: Research Track) , 2022

Bib PDF Website

@inproceedings{merkhofer2022practical,
  title = {Practical Attacks on Machine Translation Using Paraphrase},
  author = {Merkhofer, Elizabeth M. and Henderson, John and Gertner, Abigail S. and Doyle, Michael and Wong, Lily},
  booktitle = {Proceedings of the 15th biennial conference of the Association for Machine Translation in the Americas (Volume 1: Research Track)},
  pages = {227--239},
  year = {2022},
}